#!/bin/bash

# Copyright (C) 2010 Jonathan Kamens <jik@kamens.brookline.ma.us>
#
# The current version of this script is available from
# <http://stuff.mit.edu/~jik/software/bogofilter/>.
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version. 
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
# General Public License for more details. 

# Command-line options:
#
# --help	Print help message and exit
# --keep	Keep existing /tmp/notspam and /tmp/bogospam files instead of
#		regenerating them from ~/Mail/notspam* and ~/Mail/bogospam*
# --spamok	Ignore the fact that some of the messages in /tmp/bogospam were
#		classified by bogofilter as ham (i.e., bogofilter is wrong
#		about these messages).
# --hamok	Ignore the fact that some of the messages in /tmp/notspam were
#		classified by bogofilter as spam.
# --wordlist	Use the specified wordlist instead of the default.

# How to use the script:
#
# 1. Edit the "for file in" code block below as necessary to tell the script
#    where your ham and spam archives are and how they're compressed.
# 2. Run the script.
# 3. If it complains about spam messages in /tmp/notspam, load the file into an
#    editor, look for "X-Bogosity: Spam" lines, and confirm that the messages
#    that contain them are all in fact ham.  If not, then move them from your
#    ham mbox archive to your spam mbox archive and run the script again.
# 4. Ditto for ham messages in /tmp/bogospam, but this time look for
#    "X-Bogosity: Ham"
# 5. When /tmp/notspam and /tmp/bogospam are correct, run with "--hamok
#    --spamok --keep" to tell the script not to regenerate them and to go ahead
#    and do the tuning without worrying about seemingly misclassified messages.
# 6. When it's done, copy the results into your .bogofilter.cf.

USAGE="Usage: `basename $0` [--help] [--keep] [--spamok] [--hamok] [--wordlist]"

while [ "$*" ]; do
    case "$1" in
	--verbose) shift; VERBOSE=-v ;;
	--help) shift; echo $USAGE; exit ;;
	--spamok) shift; SPAMOK=yes ;;
	--hamok) shift; HAMOK=yes ;;
	--keep) shift; KEEP=yes ;;
	--wordlist) shift; WORDLIST=-D ;;
	*)  echo "Unrecognized option: $1" 1>&2
	    echo "$USAGE" 1>&2
	    exit 1
	    ;;
    esac
done

set -x

cd ~/Mail

if [ ! "$KEEP" ]; then
    for file in notspam bogospam; do
	rm -f /tmp/$file
	(bzcat $file.*.bz2; cat $file.? $file) |
        perl -e '
open(MSF, "$ENV{HOME}/.bogofilter/milter-subject-filters");

while (<MSF>) {
    s/^\s+//;
    s/\s+$//;
    next if (/^\#/);
    next if (/^$/);
    push(@sres, $_);
}

if (@sres) {
    $sre = "(?:" . join("|", @sres) . ")";
    $sre = qr/$sre/;
}

$/ = "\n\From ";

while (<>) {
    ($header, $body) = split(/\n\n/);
    if (($sre and $header =~ /\nSubject:\s*(.*\S)/ and $1 =~ /$sre/o)
#	or (($header =~ /\nFrom: SpamQuarantine\@mit\.edu/) and
#	    ($header =~ /\nSubject: Spam Quarantine Summary/))
	) {
	print "From " if ($. == 1);
    } else {
	print;
    }
}
' |
        bogofilter -M -p > /tmp/$file
    done
fi

if [ ! "$SPAMOK" ]; then
    if grep -q -s '^X-Bogosity: Ham' /tmp/bogospam; then
	echo "WARNING: Ham in /tmp/bogospam; specify --spamok (and maybe --keep) to override" 1>&2
	exit 1
    fi
fi

if [ ! "$HAMOK" ]; then
    if grep -q -s '^X-Bogosity: Spam' /tmp/notspam; then
	echo "WARNING: Spam in /tmp/notspam; specify --hamok (and maybe --keep) to override" 1>&2
	exit 1
    fi
fi

if ! bogotune $WORDLIST $VERBOSE -T 0 -n /tmp/notspam -s /tmp/bogospam; then
    if [ "$WORDLIST"x == ""x ]; then
	echo ""
	echo "Trying again with internal word list."
	echo ""
	bogotune -D $VERBOSE -T 0 -n /tmp/notspam -s /tmp/bogospam
    fi
fi

if [ $? == 0 ]; then
    rm -f /tmp/notspam /tmp/bogospam
fi